develooper Front page | perl.mvs | Postings from April 2018

Re: Smoke [blead] v5.27.10-119-g5f6af817ad FAIL(XM) os/390 25.00(2964/)

Thread Previous | Thread Next
From:
Karl Williamson
Date:
April 19, 2018 02:08
Subject:
Re: Smoke [blead] v5.27.10-119-g5f6af817ad FAIL(XM) os/390 25.00(2964/)
Message ID:
afb16e82-eae9-fc73-4b4f-895944c60224@khwilliamson.com
On 04/17/2018 01:11 AM, Yaroslav Kuzmin wrote:
> Yes . I can actually reasonably bisect. Using the git for z/OS
> 

Ok.  I have it set up in the branch

smoke-me/khw-yaroslav

That branch is different from blead because it changes the history so 
that things that cause EBCDIC not to compile were fixed up.  Otherwise 
the bisect would keep finding compilation errors, and not work properly.

To run the bisection do something like


git clone ... bisect
cd bisect
git checkout smoke-me/khw-yaroslav
/path/to/a/functioning/perl/Porting/bisect.pl --start 3b6c52ce7d --end 
716a866437e -e 'exit (uc("\x{587}") ne "\x{535}\x{552}");'

This tool has never been run AFAIK on an EBCDIC machine.  Let's cross 
our fingers that it works.  If it works, after some length of time, it 
will print a result.  To watch its progress, I pipe it to tee.

above-command | tee ../bisect.out

And then in another window, I do this

tail -f bisect.out | grep steps

I don't know how much of that will work on z/OS.

It can be tricky (at least for me) to get the command exactly right.  I 
tested the above-command on my machine (but I don't have the bug).  If I 
screwed up and it says the end commit works, you can add the

--expect-fail

parameter.  Or you can play around until it works.  The file with the 
tool contains its pod

Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About